Comments on Integrating Distance Sampling and Minimum Count Data (Schmidt et al. 2019)

Becker, Earl F.; Herreman, Jason K.

doi:10.1002/jwmg.22018

Cited by 1 publication

(1 citation statement)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By effectively separating the analyses into 2 parts, the CDS portion of the PI model remains pooling robust. Some recent work has mistakenly suggested that unmodeled heterogeneity can result in bias above and beyond that due to g (0) < 1 (Becker and Christ 2019, Becker and Herreman 2021), but this is not the case because the components of the likelihood are independent (Laake et al 2008). Even though pooling robustness breaks down when g (0) < 1 in the sense that covariates may be required to account for heterogeneity in the MR data, MCDS is not required when assuming PI.…”

Section: Detection (Pp Pa Pd)mentioning

confidence: 99%

Distance sampling surveys: using components of detection and total error to select among approaches

Schmidt

Thompson

Wilson

et al. 2022

Wildlife Monographs

View full text Add to dashboard Cite

Wildlife population estimators often require formal adjustment for imperfect detection of individuals during surveys. Conventional distance sampling (CDS) and its extensions (mark‐recapture distance sampling [MRDS], temporary emigration distance sampling [TEDS]) are popular approaches for producing unbiased estimators of wildlife abundance. However, despite extensive discussion and development of distance sampling theory in the literature, deciding which of these alternatives is most appropriate for a particular scenario can be confusing. Some of this confusion may stem from an incomplete understanding of how each approach addresses the components of the detection process. Here we describe the proper application of CDS, MRDS, and TEDS approaches and use applied examples to help clarify their differing assumptions with respect to the components of the detection process. To further aid the practitioner, we summarize the differences in a decision tree that can be used to identify cases where a more complex alternative (e.g., MRDS or TEDS) may be appropriate for a given survey application. Although the more complex approaches can account for additional sources of bias, in practical applications one also must consider estimator precision. Therefore, we also review the concept of total estimator error in the context of comparing competing methods for a given application to aid in the selection of the most appropriate distance sampling approach. Finally, we detail how the use of more advanced techniques (i.e., informed priors, open‐population distance sampling models, and integrated modeling approaches) can further reduce total estimator error by leveraging information from existing and ongoing data collection. By synthesizing the existing literature on CDS, MRDS, TEDS and their extensions, in conjunction with the concepts of total estimator error and the components of the detection process, we provide a comprehensive guide that can be used by the practitioner to more efficiently, effectively, and appropriately apply distance sampling in a variety of settings.

show abstract

Section: Detection (Pp Pa Pd)mentioning

confidence: 99%